Market‐specific Sunk Export Costs: The Impact of Learning and Spillovers

• Author: Hege Medin,Per Botolf Maurseth
• DOI: http://doi.org/10.1111/twec.12420
• Date: 01 June 2017
• Published date: 01 June 2017

Market-specific Sunk Export Costs: The

Impact of Learning and Spillovers

Per Botolf Maurseth

1

and Hege Medin

2,3

1

Department of Economics, BI –Norwegian Business School, Oslo, Norway,

2

Norwegian Institute of

International Affairs (NUPI), Oslo, Norway and

3

Norwegian School of Economics, Bergen, Norway

1. INTRODUCTION

RECENT years have seen the emergence of a literature, which incorporates sunk export

costs in models of international trade. This literature shows that, in the presence of such

costs, not all firms export (see Melitz, 2003 or also Medin, 2003 for a model wi th firms with

equal marginal production costs). Several empirical studies find evidence of sunk export costs

by analysing export persistence in firm-level data (Roberts and Tybout, 1997; Bernard and

Jensen, 2004). These studies focus on firms’ decisions of whether or not to export as such

and hence on global sunk export costs.

1

As Melitz and Redding (2014, p. 14) underline, the nature of trade costs is potentially

important but remains ‘underexplored’.

2

If sunk export costs are country and/or product speci-

fic, firms will typically serve different sets of markets, and persistence will be country and/or

product specific. Therefore, country and product-specific sunk export costs may influence

aggregate trade patterns. A few studies find evidence of country-specific sunk export costs,

but, to our knowledge, no studies have investigated the importance of sunk costs of exporting

a particular product to a particular country.

3

We refer to such costs as ‘market specific’.

We are grateful for referee comments. For comments and suggestions, we also wish to thank Ragnhild

Balsvik, Arne Melchior, Andreas Moxnes, Karen Helene Ulltveit-Moe, Fulvio Castellacci, Steinar Hol-

den, Espen R. Moen, Kjell Erik Lommerud, Leo Andreas Gr€

unfelt, Francesca Sanna-Randaccio, Jan I.

Haaland, Kjell Gunnar Salvanes and seminar participants at the international trade workshop series, UC

Berkeley, 2015; EEA annual meeting, Gothenburg, 2013; at the ETSG conference, Leuven, 2012; at the

NOITS seminar, Reykjavik, 2012; at the ESOP centre at Department of Economics, University of Oslo,

2012; at the annual Research Meeting in Economics 2012; Norwegian University of Life Science and at

BI –Norwegian Business School, 2011. Copyediting by Chris Saunders and Susan Høivik is highly

appreciated. Research funding was provided by the Research Council of Norway, project 139982/150

‘Globalization and Internationalization of the Norwegian Economy’; project 233836 ‘Traders in the Food

Value Chain: Firm Size and International Food Distribution.’ (Medin); and project 183522, ‘R&D,

Industry Dynamics, and Public Policy’ (Maurseth). The authors declared that we have no relevant mate-

rial financial interests related to the research described in this paper.

1

In the presence of such costs, temporary export promotion policies or macroshocks (such as exchange-

rate fluctuations) may have persistent effects on aggregated trade flows (Baldwin, 1988; Baldwin and

Krugman, 1989; Dixit, 1989). Generally, there is evidence of positive effects from export promotion poli-

cies (see Hiller 2012 for an overview of the literature).

2

They write: ‘The implications of different microfoundations for trade costs in models of firm

heterogeneity remain under-explored, including whether trade costs are sunk, fixed or variable’.

3

Meinen (2015) estimates the importance of country-specific sunk costs. Moxnes (2010),

Morales et al. (2011) investigate the role of country-specific versus global sunk export costs. Evi-

dence in Gullstrand (2011) suggests that country-specific sunk export costs vary with firm charac-

teristics.

©2016 John Wiley & Sons Ltd 1105

The World Economy (2017)

doi: 10.1111/twec.12420

The World Economy

Analysing only the export decision as such or the decision to export to a particular country

misrepresents sunk export costs when they are market specific.

4

The first aim of this paper was to study the importance of market-specific sunk export costs.

This is performed in a new data set of particular interest due to its high level of detail: we have

11 years of customs declaration panel data covering all Norwegian seafood exporters, the

countries they export to and the products they export. We do not therefore have to rely on sur-

vey data as do many other studies. Norway is one of the world’s largest exporters of seafood,

with an annual export value of 35.7 billion NOK in 2007 (approx. US$6.09 billion). The

industry is highly internationalised, with exports of a wide range of products to almost 200

countries. Approximately 90 per cent of all Norwegian seafood production is exported .

5

The

sector is therefore an interesting case for a study of international sales activity.

Our second aim was to study whether learning and spillovers effects lead to reductions in

market-specific export costs.

Schmeiser (2012) develops a theoretical model where learning about exporting from export

experience in other countries reduces a firm’s entry costs to a given country, denoting it

‘learning to export’. In this paper, we allow for a range of learning effects like this: intra and

intercountry as well as intra and interproduct. If this type of learning is important, it will have

consequences for export promotion policies: benefits from such policies can be larger than

expected because export promotion can boost export to other countries or of other products

than were initially targeted.

Krautheim (2012) presents a theoretical model where knowledge acquired by other exporters

in a particular destination country may spill over to potential exporters and reduce their costs of

exporting to that country. In this paper, we investigate such spillover effects in destination coun-

tries. We study spillovers both within and between products. Earlier empirical evidence is mixed

regarding spillovers that reduce global sunk export costs.

6

If, on the other hand, market-specific

spillovers are important, then policies aimed at exploiting spillovers could benefit from encour-

aging exports to certain markets rather than exports in general. Furthermore, firms targeting the

same market could benefit from organising themselves in ‘exporting societies’.

We find evidence of several different learning and spillover effects. Other recent studies

have also found indications of market or country-specific learning and spillovers, but these

studies differ from ours in the type of variables included and the econometric method applied

(see Section 5cfor an overview).

Most other studies have focused on either learning or spillovers. We include both in the

same regression, as it is conceivable that both effects could influence export costs at the same

time. We also include in the same regression discrete variables on firms’ lagged presence in

markets, capturing the extensive margin, and continuous variables on firms’ lagged export

value to markets, capturing the intensive margin. We find that learning and spillover effects

are stronger along the extensive margins than the intensive margins.

4

See Chaney (2008), Arkolakis and Muendler (2010), Bernard et al. (2011) for (static) theoretical mod-

els of country and/or product-specific sunk export costs. In the presence of such costs, only the large

and most productive firms find it profitable to export many products to many countries. Das et al. (2007)

discuss export promotion policies in the presence of fixed and sunk costs, but they do not incorporate

market-specific effects.

5

Figure based on information from the Norwegian Seafood Council.

6

See e.g. Clerides et al. (1998), Bernard and Jensen (2004) for dynamic frameworks; and Aitken et al.

(1997), Barrios et al. (2003), Greenaway et al. (2004) for static frameworks.

©2016 John Wiley & Sons Ltd

1106 P. B. MAURSETH AND H. MEDIN